A wide variety of off-line light-emitting diode (LED) drivers are known. For example, a capacitive drop off-line LED driver from On Semiconductor (Application Note AND8146/D) is a non-isolated driver with low efficiency, is limited to delivering relatively low power, and at most can deliver a constant current to the LED with no temperature compensation, no dimming arrangements, and no voltage or current protection for the LED.
Other isolated off-line LED drivers also have wide-ranging characteristics, such as a line frequency transformer and current regulator (On Semiconductor Application Note AND 8137/D); a current mode controller (On Semiconductor Application Note AND8136/D); a white LED luminary light control system (U.S. Pat. No. 6,441,558); LED driving circuitry with light intensity feedback to control output light intensity of an LED (U.S. Pat. No. 6,153,985); a non-linear light-emitting load current control (U.S. Pat. No. 6,400,102); a flyback as an LED Driver (U.S. Pat. No. 6,304,464); a power supply for an LED (U.S. Pat. No. 6,557,512); and a voltage booster for enabling the power factor controller of an LED lamp upon a low AC or DC supply (U.S. Pat. No. 6,091,614).
In general, these various LED drivers are overly complicated. Some require control methods that are complex, some are difficult to design and implement, and others require many electronic components. A large number of components can increase cost and reduce reliability. Many drivers utilize a current mode regulator with a ramp compensation in a pulse width modulation (“PWM”) circuit. Such current mode regulators require relatively many functional circuits, while nonetheless continuing to exhibit stability problems when used in the continuous current mode with a duty cycle or ratio over fifty percent. Various attempts to solve these problems utilized a constant off-time boost converter or hysteretic pulse train booster. While these solutions addressed problems of instability, these hysteretic pulse train converters exhibited other difficulties, such as elevated electromagnetic interference, inability to meet other electromagnetic compatibility requirements, and relative inefficiency. Other attempts, such as in U.S. Pat. No. 6,515,434 B1 and U.S. Pat. No. 6,747,420, provide solutions outside the original power converter stages, adding additional feedback and other circuits, rendering the LED driver even larger and more complicated.
Widespread proliferation of solid state lighting systems (semiconductor, LED-based lighting sources) has created a demand for highly efficient power converters, such as LED drivers, with high conversion ratios of input to output voltages. In order to reduce the component count, such converters may be constructed without isolation transformers by using two-stage converters with the second stage running at a very low duty cycle (equivalently referred to as a duty ratio), thereby limiting the maximum operating frequency, resulting in an increase in the size of the converter (due to the comparatively low operating frequency), and ultimately defeating the purpose of removing coupling transformers.
Various proposals to solve these problems have included use of quadratic power converters for providing a low output voltage with a wide DC conversion range, such as the quadratic power converter 10 illustrated in FIG. 1. For example, in “Switching Converter with Wide DC Conversion Range” (D. Maksimovic and S. Ćuk, May 1989 HFPC Proceedings and also in IEEE Transactions on Power Electronics, Vol. 6, No. 1, January 1991), the authors suggested using PWM converters having a single switch and featuring voltage conversion ratios with a quadratic dependence of the duty ratio. The cascaded buck and buck-boost topologies were designed and analytically synthesized for controlling the output voltage. When these circuits are used as a current source, however, they become as inadequate as conventional one-stage converters, and exhibit even more problems when used with a sinusoidal input current. For example, these circuits require a large capacitive filter following the rectified AC signal to continuously provide a steady DC output, thereby making power factor correction (“PFC”) practically impossible.
Referring to FIG. 1, the input DC voltage Vg 11 is applied to the first stage (buck-boost converter), comprising of transistor 20 (controlled by some type of controller 21), first inductor 15, capacitor 16, and diode 12. When the transistor 20 is conducting, for a linear (non-saturating) inductor 15, current is building substantially linearly in the inductor 15, while diode 12 is blocked by the reverse voltage during this portion of the cycle. When the transistor 20 is off, energy stored in the inductor 15 discharges into capacitor 16, diode 12 is forward biased and conducting during part of the off-time (discontinuous mode of operation, “DCM”) or completely during the off-time (continuous mode of operation, “CCM”), and the on-off cycle is repeated. The secondary stage is illustrated as a buck converter and comprises of the transistor 20, capacitor 18, second inductor 14, and diodes 13 and 17, with the load (illustrated as resistor 19) connected across capacitor 18. When the transistor 20 is conducting, energy from capacitor 16 is being transferred to the load 19 and output capacitor 18 via inductor 14, also charging it linearly, while diode 13 is conducting and diode 12 is blocked. When the transistor 20 is off and not conducting, diode 13 is reverse biased, and diode 17 is conducting, discharging inductor 14 into output capacitor 18. The operational process of the buck converter also may be either DCM or CCM. The transfer ratio of the converter 10 is
      -                  D        2                    1        -        D              ,where D is duty cycle or ratio, with the minus sign denoting that the polarity of the output voltage is reversed compared to the input voltage. Also, currents in transistor 20 and the output load are flowing in opposite directions, creating a difficult topology for sensing operational signals and providing corresponding feedback signals (e.g., both nodes “A” and “B” are at return potentials).
The above-referenced quadratic converter is designed to work as a voltage converter with a wide conversion ratio. Were this converter 10 to be used for current control in the output load, however, various issues may arise; for example, due to any imbalance of charges, voltages across capacitors 16 and 18 may not match, creating an excessive voltage across capacitor 16, which leads either to an over-design of the power stage or lower reliability, because this converter 10 cannot work if the voltage across capacitor 16 is greater than Vg 11. For the same reason, this converter 10 cannot be used in the AC/DC topologies requiring power factor correction.
Another proposed solution in U.S. Pat. No. 6,781,351, illustrated in FIG. 2, addressed the PFC problem, providing AC/DC cascaded power converters having high DC conversion ratios and improved AC line harmonics, with low input harmonic currents, a comparatively high power factor, and efficient operation for low voltage DC outputs. These converters, however, like the quadratic converters, have floating operational signals, which are referenced to different nodes of the power stage. Such floating operational signals make the provision of feedback signals to a controller extremely difficult, effectively requiring custom, application-specific controllers for power management.
The input 31 is an AC voltage, rectified by a bridge 32 and further filtered by a small capacitor 33. The buck-boost first stage 44 includes a blocking diode 34, which allows normal operation of the buck boost 44 at any value of input voltage (at node 45), thereby creating an opportunity to provide power factor correction if the on-time of a switch 40 is relatively constant. The second stage, a buck converter, comprises of capacitor 42, inductor 39, and diodes 38 and 41, and works substantially the same as the buck converter discussed with reference to FIG. 1. In order to prevent an uncontrollable rise of the voltage across first stage capacitor 36, the converter 30 uses additional components, a coupled inductor, and an additional diode (not illustrated), which negatively affect the economics of the converter 30. A more sophisticated control technique than PWM, also described in the patent, may address the imbalance of the capacitors' charge and prevent a high voltage at the first capacitor stage, without adding additional components to the power stage. Though the converter 30 is improved compared to the converter 10 because it can operate off line using an AC input, it still has floating operational signals, requiring excessively complicated feedback connections to PWM controller 46.
Switching power converters can have high internal voltages, such as up to hundreds or thousands of volts, for example. Since power switches, capacitors, and other components may operate at high internal voltage levels, they may be subject to voltage stress, such as an electrical force or stress across a component that potentially may cause it to fail. Further, it is desirable for a power converter to be able to function properly with a range of input voltages, such as those in use in different countries. For example, standard AC power voltages can range from a low of about 95 V in the U.S. to a high of about 264 V in Europe. As input voltage varies, switching power converters typically have held output voltage at a relatively constant level by adjusting the duty ratio. This prior art strategy, however, can cause voltage stress to increase dramatically over relevant portions of the input voltage range. Switches that are able to handle such high voltage stress may be difficult to obtain, if available at all, or if they are available, they may be expensive or have other undesirable characteristics such as a slow switching response, a low gain, or a high on-resistance, any of which may serve to reduce conversion efficiency. These voltage stress issues may cause engineers to avoid using or developing two-stage converters.
Accordingly, a need remains to provide a high conversion ratio converter to generate a controlled output current, with reduced voltage stress, and with a capability for control without overly complicated feedback mechanisms. Such a converter should be optimized to run using DC as well as AC input voltages. In addition, such a converter should provide significant power factor correction when connected to an AC line for input power. The converter should be able to function properly over a relatively wide input voltage range, while providing the desired output voltage or current, and without generating excessive internal voltages or placing components under high or excessive voltage stress. Also, it would be desirable to provide an LED driver controller for such a converter, included within a system for controlling a cascaded switching power converter, constructed and arranged for supplying power to one or a plurality of LEDs, including LEDs for high-brightness applications, while simultaneously providing an overall reduction in the size and cost of the LED driver.